This is not a client. It is a representative plant built from the five we visited in October 2026: ₹100 Cr turnover, one site, raw metal bought by the truckload, finished conductor on reels, Tally for accounts, Excel for everything else, six people in accounts and MIS. Every figure below is from published research, cited at the end, applied to that plant. When the first real engagement passes 90 days, this note is replaced by one with the client’s numbers.
What runs on Excel today
Weighbridge slips photographed on WhatsApp. A stores register. A production log per shift. Costing done quarterly on last quarter’s average metal price. Month-end close in 8 to 10 days. Nobody can say which of last month’s 40 orders lost money.
What we would deploy
- Daily brief agent. Reads Tally, the production log and the dispatch slips nightly. One message at 9 am: made, dispatched, billed, collected, cash, and the differences.
- Batch costing and yield agent. Metal in, conductor out, scrap, power and labour per run. Margin per order the morning after.
- Route and weight agent. Weighbridge at the depot, GPS en route, weighbridge at the gate. Any shortfall beyond tolerance flagged before unloading.
What the published figures say it saves
| Line | Basis | Low | Typical | High |
|---|---|---|---|---|
| Manual MIS and reconciliation hours removed | 6 people × 35% of the week on manual data work × 64% automatable × ₹23,000 a month loaded | ₹13 lakh | ₹19 lakh | ₹30 lakh |
| Raw material loss recovered | 1.9% of cost of goods median shrinkage, half recovered in year one, on ₹70 Cr of metal | ₹18 lakh | ₹66 lakh | ₹1.1 Cr |
| Freight bought by tender instead of phone | 3 to 15% of ₹4 Cr freight (Sheffi) | ₹12 lakh | ₹32 lakh | ₹60 lakh |
| Orders found to be loss-making and repriced | Not in published data; stated from client work only after go-live | - | - | - |
| Total | ₹43 lakh | ₹1.2 Cr | ₹2.0 Cr |
Month-end close moves from 8 to 10 days towards the published median of 6.4 and the top-performer figure of 4.8, which is time rather than rupees and is not counted above.
What we would put in writing
Not the typical column. The proposal would name one number from the plant’s own registers: usually the loss percentage on inbound metal, because it is measured at two weighbridges and nobody can argue with it. If it is not met in 90 days from go-live, we keep working at no charge for up to 90 more days until it is.
What we got wrong, in advance
The shrinkage benchmark is cross-industry and median; a well-run wire plant may already be below it, in which case the saving there is smaller and the costing agent matters more. The freight figure assumes freight has never been tendered; if it has, use the low end. Both are reasons the plant visit comes before the number.